(unknown)

At Tue, 26 Sep 2000 11:29:50 -0400, Simon St.Laurent wrote [1]
> At 09:59 AM 9/26/00 -0400, Mark Baker wrote:
> >My understanding from Dan was that the cat's already out of the bag; SVG
> >uses +xml.
> 
> On fragment identifiers, SVG, SMIL, and XHTML all have potential problems.
> 
> I'm even more concerned that there are a lot of XML applications out there
> which haven't considered the problem at all - it's not something that comes
> up as a question in XML tutorials that I've seen.  (I'll be adding it to
> future editions, believe me.)
> 
> On the other hand, I'm not sure that the problem is quite as drastic as it
> seems.  Any XML application can point into an SVG, SMIL, or XHTML document
> using XPointer.  It's just that SVG, SMIL, and XHTML applications may not
> let those documents point _out_ using XPointer.
> 
> While I'd like to say across the board that EVERYONE should use XPointer in
> all its glory when they use XML, it's awfully hard to push that when the
> W3C hasn't made that case across the board.
> 
> It's hard to push it while we still have language like this [1]:
> 
> >XPointer [XPTR] allows components of XML documents to be addressed 
> >in terms of their placement in the XML structure rather than on 
> >their unique identifiers. This allows referencing of any portion 
> >of an XML document without having to modify that document. Without 
> >XPointer, pointing within a document may require adding unique 
> >identifiers to it, or inserting specific elements into the document, 
> >such as a named anchor in HTML. XPointers are put within the fragment 
> >identifier part of a URI [URI]. The SMIL 2.0 specification does not 
> >require that browsers be able to process XPointers in SMIL 2.0 URI 
> >attributes. 
> 
> [1] -
> http://www.w3.org/TR/smil20/extended-linking.html#SMILLinking-Relationship-t
> o-XPointer
> 
> Simon St.Laurent
> XML Elements of Style / XML: A Primer, 2nd Ed.
> XHTML: Migrating Toward XML
> http://www.simonstl.com - XML essays and books

The above text from the SMIL spec is not meant to demean XPointer, it
is there to discourage misconceptions about its potential relation
with SMIL.  The reason that the SMIL 2.0 spec does not require that
browsers process XPointer is that browsers don't and can't process
XPointer -- servers do.  Please correct me if my understanding of the
general architecture is too primitive, but servers process XPointer,
not clients -- right?.  Clients send URI's to the given server, the
server processes it and returns the located data to the client.  Since
browsers are clients, they don't process XPointer, they just package
up much of the URI string and send it out, and wait for the response.

Keep in mind that SMIL does not require the use of any format other
than SMIL, even though it is completely dependent on other formats.
SMIL is an integration language: it contains locations of media
objects and descriptions of how these are combined to make a
multimedia presentation.  The formats of these media objects is not
specified for SMIL.  Authors in SMIL choose formats for their
integrated media that they can determine receives substantial support
in SMIL browsers.  It's a free-market integration format, and that's
the best way to go about it.

Similarly, authors will also choose locations format like XPointer in
SMIL if there there are servers that process it and return useful
media data.  Authors are completely able to use XPointer in SMIL
presentations, and nothing in the current spec inhibits this.  Some
SMIL attributes take URI's, primarily to locate media, and URI's these
days can contain XPointer.  The ability of browsers to process
presentations with XPointer depends not on their processing XPointer,
since that's done by servers, but their ability to process the media
these servers return and to integrate it into the presentation.
Authors will choose to use XPointer in SMIL presentations if there are
XPointer servers at the URI's giving media they need, and if the SMIL
browsers can processes the non-XPointer media data that is returned.

I believe what's written above applies to HTML, and perhaps to SVG as
well.

Nonetheless, XPointer is certainly a specified and required part of
SMIL.  SMIL uses XPointer values in certain attributes for internal
SMIL-to-SMIL communication, such as specifying endpoints of links and
synchronization relationships.

-Lloyd

[1] http://www.imc.org/ietf-xml-mime/mail-archive/msg00583.html

--
Lloyd Rutledge  vox: +31 20 592 41 27       fax: +31 20 592 41 99
CWI             net: [email protected]  Web: http://www.cwi.nl/~lloyd
Post:   PO Box 94079   |  NL-1090 GB Amsterdam  |  The Netherlands
Street: Kruislaan 413  |  NL-1098 SJ Amsterdam  |  The Netherlands